Aquaponics is the practice of growing plants without soil, and it combines hydroponics (growing plants in water) and recirculating aquaculture (raising fish) systems. Some of the water’s nutrients are lost during the recirculation process in an aquaculture system, but the plants in the system are able to clean the water by absorbing the nutrients. The increasing problem of food scarcity has prompted more creative approaches to urban farming. Aquaponics is a type of hydroponics that can be used to grow plants in an urban environment. However, in order to grow plants, a smart aquaponic system requires close monitoring, mechanization, and management. Implementing vision-based systems that incorporate algorithms that make use of machine learning to boost agricultural output is one promising way to put this theory into practice. To do this, we analyzed how well the approximations of Logistic Regression (LR), K-Nearest Neighbor (KNN), and Linear Support Vector Machine (L-SVM) performed. To do this, we used computer vision to extract features from images of celery grown in a computer-controlled aquaponics system and used those images as training data for the algorithms. In order to teach the algorithms, these pictures were taken. The backup systems and cross-validation checks for each method have been improved. CNN was shown to be the most successful method
Read full abstract